Generative Images - Generative Imageries: Challenges of Visual Communication (Research) in the Age of AI

• Datum: 20. November 2024 – 22. November 2024
• Street: Linzer Straße 4
• Location: ZeMKI

Societal discourse about the benefits, risks, and challenges of artificial intelligence (AI) is in full swing. To date much of the debate has focused on text-to-text generators such as Chat-GPT, Bing, or Bard (for a recent overview see Sachs-Hombach et al. 2023). Yet, rapid AI development is also fundamentally changing visual communication: With Stable Diffusion, Dall-E or Midjourney, powerful AI generators of images and image sequences have long been available whose output is impressive in terms of diversity, richness of detail and image complexity. AI generated images have long been used in a variety of media, social, and political contexts — even for strategic persuasion and political propaganda — and without the generative origin being obvious to users or being disclosed.

Generative images — generative imageries (GI) — thus offer enormous potential to fundamentally change the production, use, reception and handling of images. At the same time, little light has been shed on what it actually means for our understanding of imagery when images are increasingly produced generatively. Does this mean the “end of truth” as the established German news magazine Der Spiegel headlined in July 2023 emphasizing the challenges of generative imagery?